Purpose: With the increased availability of clinical deformable image registration tools, summing dose distributions calculated over different CT volumes is a tempting option. However, visual inspections of image registrations are insufficient to assess accuracy, and no patient‐specific QA methods exist to‐date. If the plan quality of the resulting summed dose is poor, it is difficult to discern whether the degradation results from organ motion or registration inaccuracies. We present a straightforward patient‐specific test that identifies invalid dose registrations. Methods: We applied the proposed test to assess dose registration accuracy for an illustrative 2‐fraction HDR brachytherapy prostate case. One plan per fraction was optimized using a unique CT acquired and contoured for each fraction and MIMVista was used to deformably register the clinical dose and contours. The resulting deformed and target CTs were compared. DVHs for the original dose distributions were calculated on their corresponding contours and for the deformed dose using both the target and warped contours. The four resulting DVH sets were compared and analyzed. Results: Though the deformed and target CTs match closely, V100 for the deformed dose plan is 42.0% as compared to 88.6% and 89.6% for the first and second day plans. Computing the deformed dose prostate DVH on the warped contours instead produces a target DVH with V100 of 82.6%, which differs greatly from the deformed V100 calculated on the target contours. Conclusion: When assessing a deformed dose, it is impossible to determine a priori whether degradation of the plan DVHs has resulted from organ motion or registration inaccuracies. If contoured and registered properly, the DVHs of the deformed dose should be similar regardless of whether they are computed over the deformed or target CT contours. Large differences between the two are a strong indication that the warped dose is unusable.
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging